Skip to main content
Use V6 to dress a person in up to four items at once. Send the person photo and one outfit array containing each item. The shape is the same as a single-item request.
Replace the example URLs with direct image files you control. This creates one output with three garments. Setting num_images: 2 requests two result images, each showing the same three-item outfit; each result has its own job.

Choose one reference format

Use one of these arrays per request. To combine inputs that currently have different formats, upload the URL images first or save all items as outfit assets. Each array entry represents one item. Front and back photos of the same garment belong in one asset, rather than being separate outfit items. Choose compatible layers, such as a shirt, trousers, and jacket. A prompt can clarify how to wear them, but does not guarantee exact layering or garment reproduction. Inspect each result for logos, prints, seams, colour, fit, and occluded details. The four-item limit applies to V6. Explicitly choosing an older model can reduce the supported number of items. Quality and results โ†’ ยท Full request reference โ†’